Methodology: where Fashio AI's numbers come from
Published October 12, 2026 · figures or their rules last changed October 7, 2026
Fashio AI publishes four kinds of numbers: our own product figures (credit costs, plans, library counts, output limits), a dated price table of competing tools, reference tables in our guides (size conversions, colour seasons), and cost estimates. This page says where each comes from and what it does not show. We make and sell Fashio AI, which is also one of the tools in the price table; the disclaimer lists every commercial interest. Changes to any of this are dated on the changelog.
Fashio AI's own figures
Every figure on Fashio AI by the Numbers is read from Fashio AI's production configuration (the seed files and server constants the running platform uses) and counted programmatically, not transcribed. The page was first published on 2026-07-27; the figures were last re-verified on 2026-10-06. The 101 AI models, 87 poses, 50 scene backgrounds and 195 UGC presets are full counts of that seed data, not a sample.
Credit costs come from the style table the charging code reads, and a few operations from the price constant in the module that charges for them; the page marks which is which. Plan prices and credit allowances come from the web billing catalogue. Top-up packs (40 credits $9.99, 120 credits $29.99, 1,200 credits $129.99) require an active plan.
Cost per image is arithmetic, shown in full. Plan price ÷ credits in the plan × credits per operation. Example: $29.99 ÷ 160 credits × 4 credits = $0.75 per fashion image on Plus monthly. Nothing else goes into it: no assumed retry rate, no assumed share of usable outputs.
The same figures are filled into /llms.txt and the product pages at build time from one file, so the pages cannot quote a different number. Machine-readable copy: /data/platform-numbers.json.
Competitor prices
The price table covers 15 tools: Fashio AI and 14 others. Every figure is read from the vendor's own pricing page, help centre or API documentation, never from another comparison site and never from one competitor's description of another. Each row carries the date it was read (currently 2026-02-24 to 2026-08-27; our own row 2026-10-06) and the URLs it was read from.
Where a vendor does not publish a number, the table says "not published"; it does not estimate one. A per-image cost is shown only when the vendor publishes both a plan price and a credit cost per generation, and it is plain division, the same as for our own figures. Rows older than 120 days are flagged on the page as needing a re-check. The table was first published on 2026-08-17. Machine-readable copy: /data/pricing-comparison.json.
| Coverage label | What it means |
|---|---|
| Full garment range | Puts any garment on a model. |
| Restricted | On-model output, with a category restriction the vendor states (for example, tops only). |
| No on-model output | Generates no human model at all (backgrounds, enhancement, outpainting). |
| Human photography | Real models in a real studio, booked as a service. |
How competitors are evaluated
- Price, coverage and published terms only. The table compares what each vendor publishes: entry price, what the entry tier includes, per-image cost where derivable, free tier, garment coverage, and terms such as minimum commitments or expiring credits. It does not score, rank or rate output quality.
- Head-to-head pages. 14 tools have a "Fashio AI vs" page under /alternatives. They are written by the team that makes Fashio AI, and each one states where the other tool is the better choice.
- Rows follow the vendor's own page. A row changes only when the vendor's published pricing or documentation changes, and the new read date is shown. We have no affiliate links to any vendor in the table.
- A cheap generation is not a finished asset. Per-image figures are not quality judgements; test any tool on your own garment before buying.
Size charts and the size converter
The size converter on the size-chart guides holds 16 women's and 9 men's rows. They are standard ready-to-wear conversions anchored to body measurements, not to a label.
- Women: US numeric size is the spine. UK = US + 4, EU (German) = US + 30, FR = EU + 2, IT = EU + 4. Bust, waist and hip follow Target's women's regular and plus body chart (read 2026-10-06), which matches Marks & Spencer within half an inch where the two overlap.
- Men: body chest in inches is the spine. Chest, waist and neck follow the Marks & Spencer men's general and formal-shirt guides (read 2026-10-06). EU = chest + 10, rounded to the EU step.
- India: the men's India number is the number printed on Indian shirt tags, checked against U.S. Polo Assn. India, Park Avenue, Blackberrys, Monte Carlo, Peter England, Powerlook and Indian Terrain charts on 2026-10-06. It is a label, not a chest measurement, and brands disagree by one step.
- Straight and plus sizes are kept apart. XL/XXL/XXXL and 1X/2X/3X are different scales that share letters.
- No brand guarantee. Individual brands vary by up to a full size either way. Measure, then check the brand's own chart.
Machine-readable copy: /data/size-charts.json.
Colour season analysis
The colour analysis is a 11-question questionnaire, not a photo analysis; it accepts no upload. Each answer adds weights on three axes (hue, value, chroma), the 12-season model's axes. Scores are normalised by the maximum each axis can reach, which is derived from the questions themselves. The parent season comes from the signs of hue and chroma; the sub-season is the dominant eligible axis, giving one of 12 seasons.
The classifier is deterministic: the same answers always give the same season. An axis within ±0.08 of zero counts as a tie: a hue tie is settled by the two direct undertone questions (metal, white paper), a chroma tie by the value axis. A result is labelled strong when the dominant axis leads the next by more than 0.18, likely above 0.07, and borderline otherwise.
There is no canonical public hex set for the seasons, because the commercial swatch books are not published. Each palette includes published anchor colours, and the rest are chosen inside that season's hue, value and chroma range. They are not official palettes.
Cost estimates and reports in the guides
The photoshoot cost guide gives day-rate and rental ranges and a formula for adding them up (talent × usage, photographer, studio, crew, equipment, retouching, transport, agency fee, contingency). Those ranges are our planning estimates, not survey data and not sourced line by line. Use them to sanity-check a budget, and use real quotes to set one.
The AI fashion industry report has its own methodology section: every number in it is either sourced to a named third party with a link or is our own pricing, and it states that it gives no market-size total, adoption percentage or cross-industry return-rate average because we found none we would stand behind.
What we do not claim
- A quality ranking, score or rating of any tool, including ours.
- A median generation time. We publish a 300-second server-side ceiling, which is not a typical wait.
- Customer counts, images-processed totals or review scores.
- That a size conversion fits at a specific brand, or that a colour season is an official or professional draping result.
- That a competitor price is current after the date shown on its row. Prices change; check the vendor's page before budgeting.
Corrections and change history
If a figure here is wrong, email [email protected] with the page and what you found. A confirmed correction is fixed and dated on the changelog (also as /data/changelog.json), with the figure before and after.